---
base_model: royallab/MN-LooseCannon-12B-v2
language:
- en
license: apache-2.0
tags:
- text-generation-inference
- transformers
- unsloth
- mistral
- trl
- sft
---

# Uploaded  model

- **Developed by:** Trappu
- **License:** apache-2.0
- **Finetuned from model :** royallab/MN-LooseCannon-12B-v2

# Details

This model was trained on my own little dataset free of synthetic data, which focuses solely on storywriting and scenrio prompting (Example: `[ Scenario: bla bla bla; Tags: bla bla bla ]`), 

I don't really recommend this model due to its nature and obvious flaws (rampant impersonation, stupid, etc...). It's a a one-trick pony and will be really rough for the average LLM user to handle. 

Instead, I recommend you guys use [Magnum-Picaro-0.7-v2-12b](https://huggingface.co/Trappu/Magnum-Picaro-0.7-v2-12b). The idea was to have Magnum work as some sort of stabilizer to fix the issues that emerge from the lack of multiturn/smart data in Picaro's dataset. It worked, I think. I enjoy the outputs and it's smart enough to work with.

# Prompting

If for some reason, you still want to try this model over Magnum-Picaro, it was trained on chatml with no system prompts, so below is the recommended prompt formatting.

```
<|im_start|>user
bla bla bla<|im_end|>
<|im_start|>assistant
bla bla bla you!<|im_end|>
```

For SillyTavern users:

[Instruct template](https://firebasestorage.googleapis.com/v0/b/koios-academy.appspot.com/o/trappu%2FChatML%20custom%20Instruct%20template.json?alt=media&token=9142757f-811c-460c-ad0e-d04951b1687f)

[Context template](https://firebasestorage.googleapis.com/v0/b/koios-academy.appspot.com/o/trappu%2FChatML%20custom%20context%20template.json?alt=media&token=0926fc67-fa9f-4c86-ad16-8c7c4c8e0b64)

[Settings preset](https://firebasestorage.googleapis.com/v0/b/koios-academy.appspot.com/o/trappu%2FHigh%20temp%20-%20Min%20P%20(4).json?alt=media&token=ac569562-af11-4da1-83c1-d86b25bb4fe1)

The above settings are the ones I recommend. 

Temp = 1.2

Min P = 0.1

DRY Rep Pen: Multiplier = 0.8, Base = 1.75, Allowed Length = 2, Penalty Range = 1024

Little guide on useful samplers and how to import settings presets and instruct/context templates and other stuff people might find useful [here](https://rentry.co/PygmalionFAQ#q-what-are-the-best-settings-for-rpadventurenarrationchatting)

Every other sampler neutralized.